Using REW's new inversion feature for room correction

[ppataki]

I am posting this with the kind permission of @Serkan who drew my attention to this brand new method for room correction

In this video you will see how to try this with an early access version of REW, available here:

V5.20.10 Early Access

V5.20.10/11 Early Access release There are more changes in V5.20.10 than is usually the case for the minor releases, so I am providing an early access version here for feedback ahead of the formal release. Windows 64-bit Windows 32-bit MacOS Linux AMD64 Linux no JRE (for non-AMD64 linux...

www.avnirvana.com

I have tried this new method and personally I like the results (both from an objective and from a subjective point-of-view) better than any EQ-ing method that I was using before

Please use this thread to share your experience and also if you have any suggestions re. fine-tuning this great new technique

Thank you

 

[ernestcarl]

I am posting this with the kind permission of @Serkan who drew my attention to this brand new method for room correction

In this video you will see how to try this with an early access version of REW, available here:

V5.20.10 Early Access

V5.20.10/11 Early Access release There are more changes in V5.20.10 than is usually the case for the minor releases, so I am providing an early access version here for feedback ahead of the formal release. Windows 64-bit Windows 32-bit MacOS Linux AMD64 Linux no JRE (for non-AMD64 linux...

www.avnirvana.com

I have tried this new method and personally I like the results (both from an objective and from a subjective point-of-view) better than any EQ-ing method that I was using before

Please use this thread to share your experience and also if you have any suggestions re. fine-tuning this great new technique

Thank you

I only had time to watch the video yesterday, but below are the results of some in-room measurements of my (right speaker) Presonus Sceptre S8 at the couch MLP:


*I mistakenly posted the Left Speaker RMS avg MMM (orange curve), but the corrected Right speaker measures similarly enough!

Thank you to @Serkan for the illustrative video run-through!


I only have few comments for now:

1. Windowing and smoothing options are quite limited
2. Doesn't currently seem to work well with simple multi-point RMS average and MMM (moving-microphone spatial average) measurements.
3. There is no available control for the number of taps and impulse centering along with a quick way to simulate an overlay graph to see "slippage" or ripple effects in the response (if attempting to economize filter generation).

In the meantime, some may appreciate reviewing the slideshow from @j_j (one of the presenters): http://www.aes-media.org/sections/pnw/pnwrecaps/2008/jj_jan08/

But the ff. are three slides that specifically highlight my concern with this simplistic inversion methodology:

 

[thewas]

I am posting this with the kind permission of @Serkan who drew my attention to this brand new method for room correction

In this video you will see how to try this with an early access version of REW, available here:

V5.20.10 Early Access

V5.20.10/11 Early Access release There are more changes in V5.20.10 than is usually the case for the minor releases, so I am providing an early access version here for feedback ahead of the formal release. Windows 64-bit Windows 32-bit MacOS Linux AMD64 Linux no JRE (for non-AMD64 linux...

www.avnirvana.com

I have tried this new method and personally I like the results (both from an objective and from a subjective point-of-view) better than any EQ-ing method that I was using before

Please use this thread to share your experience and also if you have any suggestions re. fine-tuning this great new technique

Thank you

Thanks for sharing here, didn't know about it before, looks very similar to the way of standard FIR filter generation by Acourate which I also own since a decade but personally am back to more minimalist IIR filters and mainly below transition frequency as those sound better to me, will be interesting though to hear more other experiences.

 

[ppataki]

+ kind of like this old one too if I am not mistaken:

I will post my measurements using this new method sometime this week

 

[ernestcarl]

I think FIR DSP is incredibly powerful, but the method of application or ways to go about it must be considered. While, it can turn what might first appear like a "broken" speaker design into something really good:

Fulcrum Acoustic RM22

The bulk of the finer complex magnitude and phase FIR EQ correction is applied above 500Hz in the coax driver. Only a very few simple minimum phase PEQs is needed for the lower half, and the rest above (general HF shelving filter etc.) for this specific speaker. Post installation at the venue, some more manual adjustments could be added to the DSP capable amp-processor.

It can't really truly "fix" gross directivity issues and sub-optimal room acoustics. All things considered, of course, we might be able to create something altogether different sounding (from the original out-of-box sound) and perhaps simulate other improvements via FIR filters in-room, but there must be practical limits in the real world -- at least that's what I think.

 

[ppataki]

All things considered, of course, we might be able to create something altogether different sounding (from the original out-of-box sound)

Totally agree, that is what I am doing with my MCLAs......
(and that is what you would do with any other 1-way systems for sure)

For such systems the goal really is to fundamentally and radically change/improve the frequency response and not just to apply smaller fixes
And I personally have never found the 'traditional' EQ-ing drastic enough to be able to fully correct such systems - hence I turned my attention to alternative solutions

 

[ernestcarl]

Totally agree, that is what I am doing with my MCLAs......
(and that is what you would do with any other 1-way systems for sure)

For such systems the goal really is to fundamentally and radically change/improve the frequency response and not just to apply smaller fixes
And I personally have never found the 'traditional' EQ-ing drastic enough to be able to fully correct such systems - hence I turned my attention to alternative solutions

I'd like to see more speakers that have been radically altered with EQ from others. My own corrections are only "minor", and maybe some irrelevant wiggles here and there -- but still small enough to fit within REW's generic 20 PEQ entry limit. I've heard others complain of that "small" number saying it's sorely lacking and simply not enough. I'm interested what these folks have to say, and would love it if they can also post measurements here.

 

[ernestcarl]

There is no available control for the number of taps and impulse centering along with a quick way to simulate an overlay graph to see "slippage" or ripple effects in the response (if attempting to economize filter generation).

Apologies, my bad! I realize no "economizing" is necessary at all (at the moment) with this particular method in REW as it uses purely minimum phase FIR filtering. So no time correction imposed delays.

 

[ppataki]

Here are some measurements after spending some time with the inversion method

Let's just stick to my left channel to make it easier

This is the FR (1/12 smoothing) with zero DSP:

This is the FR with some shelves applied (so the correction won't be that huge with whatever method I use)

I have used auto target setting and 8% regularization at the 1/A operation and got this (1/12 and FDW35)
(note that a high shelf was applied on purpose)

Wavelet looks like this:

Step response:

Now let's change the regularization to 0.1% (to compensate more)

FR:

Now we are talking....

Wavelet:

Notice how the curve got 'symmetrialized' - pre-ringing increased (definitely not audible) and post-ringing decreased (that is what I like)

the same is visible on the SR:

To be honest, when listening to it, both versions sound great; with the 8% one I can hear more bass due to the elevated post-ringing while with the 0.1% version bass sounds 'less' but tighter

Now I can confidently say that I do prefer this method to traditional EQ-ing or other DRC options
I can create my filters based on one measurement and with like 20-30 clicks altogether - quickly delivering awesome results!
+ multichannel works too! (very important for me)

Note that I will create another thread today or tomorrow to showcase the usage of different convolution solutions (the above measurements were done using Jriver's Convolution engine)

Any questions/comments please shoot!

 

[ernestcarl]

Here are some measurements after spending some time with the inversion method

Let's just stick to my left channel to make it easier

This is the FR (1/12 smoothing) with zero DSP:

View attachment 224359

This is the FR with some shelves applied (so the correction won't be that huge with whatever method I use)

View attachment 224361

I have used auto target setting and 8% regularization at the 1/A operation and got this (1/12 and FDW35)
(note that a high shelf was applied on purpose)

View attachment 224363

Wavelet looks like this:

View attachment 224364

Step response:

View attachment 224365


Now let's change the regularization to 0.1% (to compensate more)

FR:

View attachment 224366

Now we are talking....

Wavelet:

View attachment 224367

Notice how the curve got 'symmetrialized' - pre-ringing increased (definitely not audible) and post-ringing decreased (that is what I like)

the same is visible on the SR:

View attachment 224368

To be honest, when listening to it, both versions sound great; with the 8% one I can hear more bass due to the elevated post-ringing while with the 0.1% version bass sounds 'less' but tighter

Now I can confidently say that I do prefer this method to traditional EQ-ing or other DRC options
I can create my filters based on one measurement and with like 20-30 clicks altogether - quickly delivering awesome results!
+ multichannel works too! (very important for me)

Note that I will create another thread today or tomorrow to showcase the usage of different convolution solutions (the above measurements were done using Jriver's Convolution engine)

Any questions/comments please shoot!

Your application of 1/12 smoothing (instead of none) should help soften any excessive comb-filtering like correction above 500 Hz or so even when a longer window is applied... so that's better.

I don't think I would ever go back to a single-point measurement correction technique just out of convenience, though. More information about the room's response at multiple points (pseudo-random distribution) is always an asset for assessment purposes even if you don't choose to apply corrections from all of that data.

 

[ernestcarl]

It looks like roughly 15 cycle windowing and 1/6 smoothing filter settings gives perhaps enough "softness" in the correction filters I would prefer to see.

*note the application of limits during inversion

1/6 smoothing should also be kept during the inversion step so that the final, unsmoothed filter curve to be exported retains the same quality.

Also, more data from multi-point or spatial averaging techniques would give out milder/smoother filter curves. The excessive cuts applied around 300 Hz (from SBIR) for instance should have only been about half of what was applied. Move elsewhere in the couch or in the room and there is no peak at all at 300 Hz. Single-point data is blind to such things. There's still quite a potential for other improvements here and there, but this new feature is definitely a worthy additional tool to have in REW...

 

[fluid]

I mentioned in a previous thread that there a few things in the video that did seem right to me. One of the biggest for me was the description of the Harman Target Curve and it being the only curve anyone would need. I'll attach some words from Floyd Toole that appear in the same publication the graph of curves come from. The whole paper is free and worth reading if not already. The Measurement and Calibration of Sound Reproducing Systems.

https://www.aes.org/tmpFiles/elib/20220817/17839.pdf

"Research by Olive et al. [48] was distinctive in that the loudspeaker used was anechoically characterized, the room described [49], and high-resolution room curves measured. In the double-blind tests, listeners made bass and treble balance adjustments to a loudspeaker that had been equalized to a flat smooth steady-state room curve. The loudspeaker had previously received high ratings in independent double blind comparison tests, without equalization. Three tests were done, with the bass or treble adjusted separately with the other parameter randomly fixed, and a test in which both controls were available, starting from random settings. It was a classic method-of-adjustment experiment. For each program selection, listeners made adjustments to yield the most preferred result. In Fig. 14 the author has modified the original data to separately show the result of evaluations by trained and untrained listeners. This is compared to the small room prediction from Fig. 13(a). The “all listeners” average curve is close to the predicted target, except at low frequencies where it is apparent that the strongly expressed preferences of inexperienced listeners significantly elevated the average curve. In fact, the target variations at both ends of the spectrum are substantial, with untrained listeners simply choosing “more of everything.” An unanswered question is whether this was related to overall loudness—more research is needed. However, most of us have seen evidence of such more-bass, more-treble listener preferences in the “as found” tone control settings in numerous rental and loaner cars. More data would be enlightening, but this amount is sufficient to indicate that a single target curve is not likely to satisfy all listeners. Add to this the program variations created by the “circle of confusion” and there is a strong argument for incorporating easily accessible bass and treble tone controls in playback equipment. The first task for such controls would be to allow users to optimize the spectral balance of their loudspeakers in their rooms, and, on an ongoing basis, to compensate for spectral imbalances as they appear in movies and music."

There is also a slide from one of Sean Olive's documents that shows some more information
https://www.juloaudio.sk/Umiestnenie_reprosustav/History of Harman Target Curve.pdf



So while there is a lot of similarity in preferred room curves the idea that you can pick one and apply to all speakers with all directivities in all rooms and have them sound the same or right is not a good one.

In my own correction routine I have moved away from having DRC use a target curve to define the overall steady state response. While I have had good results using that method it can be quite time consuming and complicated making different target curves trying to find improvements.

What I do now is use DRC to correct to a flat response and then apply a layer of PEQ over the top to set the tonal balance which was judged by ear. This PEQ is a collection of shelving filters spaced an octave apart which allows the slope to be changed consistently or varied at certain points based on how it sounds. Each speaker and room would be different and this sort o approach allows easy modification of the tonal balance while listening in a controlled way. Sometimes a small change to the Q of the shelf or gain by a a small amount can impact the perception quite a lot. Eventually things start to sound more "right" for want of a better word.

Experimentation in this sort of processing is important because there is still a fair bit of alchemy in getting the best result as the method itself is impacting so many different factors at once. The usual factors pointed out as to why "room correction" is not a good idea are valid in of themselves but in my own experience if they are considered and managed then it is still possible to get a really result but one with a one size fits all curve.

 

[ppataki]

I mentioned in a previous thread that there a few things in the video that did seem right to me. One of the biggest for me was the description of the Harman Target Curve and it being the only curve anyone would need. I'll attach some words from Floyd Toole that appear in the same publication the graph of curves come from. The whole paper is free and worth reading if not already. The Measurement and Calibration of Sound Reproducing Systems.

https://www.aes.org/tmpFiles/elib/20220817/17839.pdf

"Research by Olive et al. [48] was distinctive in that the loudspeaker used was anechoically characterized, the room described [49], and high-resolution room curves measured. In the double-blind tests, listeners made bass and treble balance adjustments to a loudspeaker that had been equalized to a flat smooth steady-state room curve. The loudspeaker had previously received high ratings in independent double blind comparison tests, without equalization. Three tests were done, with the bass or treble adjusted separately with the other parameter randomly fixed, and a test in which both controls were available, starting from random settings. It was a classic method-of-adjustment experiment. For each program selection, listeners made adjustments to yield the most preferred result. In Fig. 14 the author has modified the original data to separately show the result of evaluations by trained and untrained listeners. This is compared to the small room prediction from Fig. 13(a). The “all listeners” average curve is close to the predicted target, except at low frequencies where it is apparent that the strongly expressed preferences of inexperienced listeners significantly elevated the average curve. In fact, the target variations at both ends of the spectrum are substantial, with untrained listeners simply choosing “more of everything.” An unanswered question is whether this was related to overall loudness—more research is needed. However, most of us have seen evidence of such more-bass, more-treble listener preferences in the “as found” tone control settings in numerous rental and loaner cars. More data would be enlightening, but this amount is sufficient to indicate that a single target curve is not likely to satisfy all listeners. Add to this the program variations created by the “circle of confusion” and there is a strong argument for incorporating easily accessible bass and treble tone controls in playback equipment. The first task for such controls would be to allow users to optimize the spectral balance of their loudspeakers in their rooms, and, on an ongoing basis, to compensate for spectral imbalances as they appear in movies and music."

There is also a slide from one of Sean Olive's documents that shows some more information
https://www.juloaudio.sk/Umiestnenie_reprosustav/History of Harman Target Curve.pdf

View attachment 224853

So while there is a lot of similarity in preferred room curves the idea that you can pick one and apply to all speakers with all directivities in all rooms and have them sound the same or right is not a good one.

In my own correction routine I have moved away from having DRC use a target curve to define the overall steady state response. While I have had good results using that method it can be quite time consuming and complicated making different target curves trying to find improvements.

What I do now is use DRC to correct to a flat response and then apply a layer of PEQ over the top to set the tonal balance which was judged by ear. This PEQ is a collection of shelving filters spaced an octave apart which allows the slope to be changed consistently or varied at certain points based on how it sounds. Each speaker and room would be different and this sort o approach allows easy modification of the tonal balance while listening in a controlled way. Sometimes a small change to the Q of the shelf or gain by a a small amount can impact the perception quite a lot. Eventually things start to sound more "right" for want of a better word.

Experimentation in this sort of processing is important because there is still a fair bit of alchemy in getting the best result as the method itself is impacting so many different factors at once. The usual factors pointed out as to why "room correction" is not a good idea are valid in of themselves but in my own experience if they are considered and managed then it is still possible to get a really result but one with a one size fits all curve.

I fully agree!
I use the same method: correcting to flat and then applying shelves on the top of that

 

[thewas]

What I do now is use DRC to correct to a flat response

Anechoically flat or at the LP without gating?

 

[fluid]

Anechoically flat or at the LP without gating?

Neither, at the listening position using variable frequency dependent windowing.
I have only spent any time using this method with full range line arrays which are quite different and I found they did not respond at all well to anechoic EQ.

In a previous room I used multiple positions in a vector average and that had my preference there. That room was untreated and was quite reverberant. In the current room that is deader the single position works out best.

I have tried every measurement and equalization strategy I could think of and decided what works for me with my speakers in my room.

 

[thewas]

I have tried every measurement and equalization strategy I could think of and decided what works for me with my speakers in my room.

Same here and I my personal experience is that there is no single approach that works optimally for different cases, often it is a compromise between smooth anechoic response and smooth response at the listeners position.

 

[abdo123]

Neither, at the listening position using variable frequency dependent windowing.

Wait how do you do that?

 

[abdo123]

I have only spent any time using this method with full range line arrays which are quite different and I found they did not respond at all well to anechoic EQ.

If a line array extends from floor to ceiling then you never trully leave the nearfield which makes it difficult to project current (farfield based) research on it.

 

[fluid]

Wait how do you do that?

DRC-FIR, the windows are set with a value at the lower and upper end of correction with an exponent to vary how the curve between the two is drawn. The result is quite different to a consistent fractional octave or cycle based window across all frequencies.

If a line array extends from floor to ceiling then you never trully leave the nearfield which makes it difficult to project current (farfield based) research on it.

Indeed, some factors remain constant but there is enough difference for it to not be completely valid.

 

[abdo123]

DRC-FIR, the windows are set with a value at the lower and upper end of correction with an exponent to vary how the curve between the two is drawn. The result is quite different to a consistent fractional octave or cycle based window across all frequencies.

Is this rooted in some psychacoustic research based on the prescedence effect or echo thresholds?